This repo contains the notebook Active-Learning.ipynb
where you can find different examples of using the
modAL Python package for
Active Learning.
In particular, the notebook shows how Active Learning allows to achieve better performance with fewer training samples. The difference is even more impressive for more complex Machine Learning tasks, such as use cases involving highly dimensional input spaces and multiclass classification problems.
The code was tested with Python 3.9 on Mac OS X and Linux. To make sure you have the right dependencies, please install the requirements as follows:
pip install -r requirements.txt